Method for fast and maximal data recovery in failure-prone sensor networks This technology is a method of forming a linear combination of sensor data to decrease the risk of information loss in sensor networks due to sensor network failures.
Sensor networks are extremely useful in catastrophic or emergency scenarios such as floods, fires, or earthquakes. However, current approaches in sensor networks are restricted by the limited amount of data storage available at each node in the network. Fluctuations in the rate of data generated can significantly delay the reading of data, making sensor-based networks inefficient. Additionally, sensor nodes used to collect data may fail suddenly and unpredictably, resulting in the loss of valuable data.
This technology is a method and system for forming linear combinations of data in sensor networks so that data is more likely to reach the data sink, even as nodes fail. By replicating data compactly at neighboring nodes using growth codes that increase in efficiency as data accumulates at the sink, this technology is able to increase the persistence of sensed data without initial configuration of the network. The increased speed and efficiency make this technology an improved method for sensing information in disaster-prone areas.
This technology was found to preserve about 10-20% more data, compared to when no coding was done, when tested using simulations in typical sensor topologies.
Patent Issued (US 8,655,839)
IR M06-041
Licensing Contact: Greg Maskel